This series of files compile analyses done for the global analysis of Chapter 1 (version of May 15th).

All analyses have been done with PRIMER-e 6 and R 3.6.3.

Click on the table of contents in the left margin to assess a specific analysis.
Click on a figure to zoom it


We used data from subtidal ecosystems (see metadata files for more information). Only stations that have been sampled both for abiotic parameters and benthic species were included.

Selected variables for the analyses:


1. Data manipulation

For the following analyses, independant variables are habitat parameters and heavy metal concentrations, dependant variables are diversity indices. Variables have been standardized by mean and standard-deviation.

1.1. Identification of outliers

To identify stations that are not consistent with the others, we used the multivariate Cook’s Distance (CD) on the uncorrelated variables. A significative threshold of 4 times the mean of CD has been established.

0.5 mm community

We identified the following stations as general outliers:

  • stations 30, 127, 138, 144, 183, 228, 237 for habitat
  • stations 1, 11, 22, 25, 35, 127, 132, 139, 231 for metals

They have been deleted for the following analyses.

1 mm community

We identified the following stations as general outliers:

-stations 72, 82, 107, 129, 144, 202, 249 for habitat - stations 106, 108, 110, 120, 127, 130, 139, 154, 232 for metals

They have been deleted for the following analyses.

1.2. Correlations between parameters

Correlations have been calculated with Spearman’s rank coefficient.

0.5 mm community

According to these results, the following variables are highly correlated (\(|\rho|\) > 0.80) so they have been considered together in the regressions:

  • chromium, iron and manganese (iron and manganese deleted)
  • copper, lead and zinc (copper and lead deleted)

We decided to exclude silt content, as it tends to drasticaly increase VIFs due to a marginal correlation with organic matter (\(R^{2}_{adj}\) = 0.21).

Correlation coefficients between habitat parameters (0.5 mm community subset)
  depth om gravel sand silt clay
depth 1 0.298 -0.26 0.209 0.497 -0.483
om 0.298 1 -0.504 -0.407 0.479 0.01
gravel -0.26 -0.504 1 0.019 -0.347 0.162
sand 0.209 -0.407 0.019 1 -0.007 -0.749
silt 0.497 0.479 -0.347 -0.007 1 -0.534
clay -0.483 0.01 0.162 -0.749 -0.534 1
Correlation coefficients between metals (0.5 mm community subset)
  arsenic cadmium chromium copper iron manganese mercury lead zinc
arsenic 1 0.732 0.631 0.713 0.405 0.62 0.711 0.865 0.808
cadmium 0.732 1 0.784 0.691 0.514 0.669 0.66 0.854 0.838
chromium 0.631 0.784 1 0.738 0.8 0.891 0.467 0.75 0.792
copper 0.713 0.691 0.738 1 0.571 0.738 0.612 0.843 0.928
iron 0.405 0.514 0.8 0.571 1 0.83 0.187 0.458 0.571
manganese 0.62 0.669 0.891 0.738 0.83 1 0.463 0.681 0.738
mercury 0.711 0.66 0.467 0.612 0.187 0.463 1 0.787 0.683
lead 0.865 0.854 0.75 0.843 0.458 0.681 0.787 1 0.928
zinc 0.808 0.838 0.792 0.928 0.571 0.738 0.683 0.928 1

1 mm community

According to these results, the following variables are highly correlated (\(|\rho|\) > 0.80) so they have been considered together in the regressions:

  • om and silt (silt deleted)
  • chromium, iron and manganese (iron and manganese deleted)
  • copper, lead and zinc (copper and lead deleted)
Correlation coefficients between habitat parameters (1 mm community subset)
  depth om gravel sand silt clay
depth 1 0.442 -0.026 -0.328 0.317 -0.097
om 0.442 1 -0.304 -0.798 0.841 -0.125
gravel -0.026 -0.304 1 0.124 -0.383 -0.023
sand -0.328 -0.798 0.124 1 -0.927 -0.12
silt 0.317 0.841 -0.383 -0.927 1 0.069
clay -0.097 -0.125 -0.023 -0.12 0.069 1
Correlation coefficients between metals (1 mm community subset)
  arsenic cadmium chromium copper iron manganese mercury lead zinc
arsenic 1 0.743 0.79 0.809 0.636 0.707 0.702 0.892 0.88
cadmium 0.743 1 0.772 0.639 0.541 0.662 0.669 0.833 0.812
chromium 0.79 0.772 1 0.858 0.823 0.902 0.667 0.837 0.892
copper 0.809 0.639 0.858 1 0.77 0.792 0.708 0.857 0.946
iron 0.636 0.541 0.823 0.77 1 0.869 0.412 0.595 0.753
manganese 0.707 0.662 0.902 0.792 0.869 1 0.573 0.705 0.79
mercury 0.702 0.669 0.667 0.708 0.412 0.573 1 0.844 0.743
lead 0.892 0.833 0.837 0.857 0.595 0.705 0.844 1 0.928
zinc 0.88 0.812 0.892 0.946 0.753 0.79 0.743 0.928 1

2. Permutational Analyses of Variance

Results of univariate PermANOVAs on parameters and multivariate PermANOVA on the whole benthic community are presented in the table below. Variables have been standardized by mean and standard-deviation, and abundances were (log+1) transformed.

To be added.

3. Similarity and characteristic species

To be added.

4. Univariate regressions

We used linear models for the all regressions on diversity indices. Outliers and correlated variables were removed from these analyses. Variables have been standardized by mean and standard-deviation (coefficients need to be back-transformed to be used in predictive models).

4.1. Best model selection

This step was not used here as each models are necessary.

4.2. Significative variables selection

We identified which variables were selected after an AIC procedure to predict the best the parameters. Results of the variable selection, according to AIC, are shown on the tables below:

  • for the 0.5 mm community:
Variable (or combination) S N H J
depth + - + +
om/silt + + -
gravel + + +
sand + + -
clay + + + -
Adjusted \(R^{2}\) 0.33 0.5 0.27 0.16
Variable (or combination) S N H J
arsenic - -
cadmium - -
chromium/iron/manganese + - -
mercury - -
lead/copper/zinc + +
Adjusted \(R^{2}\) 0.18 0.49 0.07 0.02
  • for the 1 mm community:
Variable (or combination) S N H J
depth + - + +
om/silt
gravel -
sand - - -
clay - - -
Adjusted \(R^{2}\) 0.25 0.03 0.34 0.1
Variable (or combination) S N H J
arsenic
cadmium - -
chromium/iron/manganese
mercury
lead/copper/zinc + +
Adjusted \(R^{2}\) 0.08 0 0.05 0

Details of the regressions, with diagnostics and cross-validation, are summarized below.

0.5 mm community
Richness/habitat
## FULL MODEL
## Adjusted R2 is: 0.33
Fitting linear model: S ~ depth + om + gravel + sand + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.02378 0.06373 0.3732 0.7095
depth 0.2425 0.07287 3.328 0.001097 * *
om 0.2862 0.07987 3.584 0.0004548 * * *
gravel 0.2016 0.09011 2.237 0.0267 *
sand 0.2542 0.1164 2.183 0.03054 *
clay 0.7458 0.1126 6.623 5.636e-10 * * *
## RMSE from cross-validation: 0.8090011
Variance Inflation Factors
  depth om gravel sand clay
VIF 1.14 1.25 1.05 1.84 1.79

## REDUCED MODEL
## Adjusted R2 is: 0.33
Fitting linear model: S ~ depth + om + gravel + sand + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.02378 0.06373 0.3732 0.7095
depth 0.2425 0.07287 3.328 0.001097 * *
om 0.2862 0.07987 3.584 0.0004548 * * *
gravel 0.2016 0.09011 2.237 0.0267 *
sand 0.2542 0.1164 2.183 0.03054 *
clay 0.7458 0.1126 6.623 5.636e-10 * * *
## RMSE from cross-validation: 0.8090011
Variance Inflation Factors
  depth om gravel sand clay
VIF 1.14 1.25 1.05 1.84 1.79

Density/habitat
## FULL MODEL
## Adjusted R2 is: 0.5
Fitting linear model: N ~ depth + om + gravel + sand + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.01472 0.05775 0.2548 0.7992
depth -0.09499 0.06604 -1.438 0.1524
om 0.5369 0.07238 7.418 7.649e-12 * * *
gravel 0.1221 0.08167 1.495 0.137
sand 0.5185 0.1055 4.914 2.27e-06 * * *
clay 0.8915 0.1021 8.736 3.96e-15 * * *
## RMSE from cross-validation: 0.7850903
Variance Inflation Factors
  depth om gravel sand clay
VIF 1.14 1.25 1.05 1.84 1.79

## REDUCED MODEL
## Adjusted R2 is: 0.5
Fitting linear model: N ~ depth + om + gravel + sand + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.01472 0.05775 0.2548 0.7992
depth -0.09499 0.06604 -1.438 0.1524
om 0.5369 0.07238 7.418 7.649e-12 * * *
gravel 0.1221 0.08167 1.495 0.137
sand 0.5185 0.1055 4.914 2.27e-06 * * *
clay 0.8915 0.1021 8.736 3.96e-15 * * *
## RMSE from cross-validation: 0.7850903
Variance Inflation Factors
  depth om gravel sand clay
VIF 1.14 1.25 1.05 1.84 1.79

Diversity/habitat
## FULL MODEL
## Adjusted R2 is: 0.26
Fitting linear model: H ~ depth + om + gravel + sand + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.05663 0.06356 0.891 0.3743
depth 0.5219 0.07268 7.181 2.822e-11 * * *
om 0.01009 0.07966 0.1267 0.8993
gravel 0.156 0.08988 1.735 0.08467
sand -0.07361 0.1161 -0.6339 0.5271
clay 0.2279 0.1123 2.029 0.04422 *
## RMSE from cross-validation: 0.8101494
Variance Inflation Factors
  depth om gravel sand clay
VIF 1.14 1.25 1.05 1.84 1.79

## REDUCED MODEL
## Adjusted R2 is: 0.27
Fitting linear model: H ~ depth + gravel + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.05587 0.06327 0.883 0.3786
depth 0.535 0.07073 7.565 3.221e-12 * * *
gravel 0.1584 0.08587 1.844 0.06707
clay 0.2884 0.06988 4.127 5.989e-05 * * *
## RMSE from cross-validation: 0.8019987
Variance Inflation Factors
  depth gravel clay
VIF 1.11 1 1.12

Evenness/habitat
## FULL MODEL
## Adjusted R2 is: 0.15
Fitting linear model: J ~ depth + om + gravel + sand + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.05235 0.06563 0.7977 0.4263
depth 0.3094 0.07505 4.122 6.142e-05 * * *
om -0.1443 0.08226 -1.754 0.08146
gravel -0.004248 0.09281 -0.04577 0.9636
sand -0.2053 0.1199 -1.712 0.08889
clay -0.2228 0.116 -1.921 0.05654
## RMSE from cross-validation: 0.8333639
Variance Inflation Factors
  depth om gravel sand clay
VIF 1.14 1.25 1.05 1.84 1.79

## REDUCED MODEL
## Adjusted R2 is: 0.16
Fitting linear model: J ~ depth + om + sand + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.05272 0.06493 0.8119 0.4181
depth 0.3095 0.07478 4.138 5.741e-05 * * *
om -0.1432 0.07888 -1.816 0.07132
sand -0.2041 0.1168 -1.748 0.08245
clay -0.2218 0.1134 -1.956 0.05228
## RMSE from cross-validation: 0.8267531
Variance Inflation Factors
  depth om sand clay
VIF 1.14 1.21 1.8 1.75

Richness/metals
## FULL MODEL
## Adjusted R2 is: 0.17
Fitting linear model: S ~ arsenic + cadmium + chromium + mercury + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.06047 0.0758 -0.7978 0.4264
arsenic -0.4048 0.1342 -3.017 0.003051 * *
cadmium -0.6834 0.1672 -4.087 7.426e-05 * * *
chromium -0.01251 0.138 -0.09062 0.9279
mercury -0.4632 0.1701 -2.723 0.007325 * *
lead 1.038 0.2052 5.06 1.333e-06 * * *
## RMSE from cross-validation: 0.8958299
Variance Inflation Factors
  arsenic cadmium chromium mercury lead
VIF 1.56 2.08 1.75 1.43 2.7

## REDUCED MODEL
## Adjusted R2 is: 0.18
Fitting linear model: S ~ arsenic + cadmium + mercury + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.05982 0.07518 -0.7956 0.4276
arsenic -0.4045 0.1337 -3.026 0.002957 * *
cadmium -0.6892 0.154 -4.474 1.6e-05 * * *
mercury -0.4584 0.161 -2.847 0.005086 * *
lead 1.031 0.19 5.427 2.524e-07 * * *
## RMSE from cross-validation: 0.8902463
Variance Inflation Factors
  arsenic cadmium mercury lead
VIF 1.56 1.92 1.36 2.51

Density/metals
## FULL MODEL
## Adjusted R2 is: 0.49
Fitting linear model: N ~ arsenic + cadmium + chromium + mercury + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.1098 0.05881 -1.867 0.06407
arsenic -0.5128 0.1041 -4.926 2.398e-06 * * *
cadmium -0.7432 0.1297 -5.729 6.212e-08 * * *
chromium 0.2695 0.1071 2.517 0.01301 *
mercury -0.7299 0.132 -5.529 1.588e-07 * * *
lead 1.43 0.1592 8.986 1.907e-15 * * *
## RMSE from cross-validation: 0.7065136
Variance Inflation Factors
  arsenic cadmium chromium mercury lead
VIF 1.56 2.08 1.75 1.43 2.7

## REDUCED MODEL
## Adjusted R2 is: 0.49
Fitting linear model: N ~ arsenic + cadmium + chromium + mercury + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.1098 0.05881 -1.867 0.06407
arsenic -0.5128 0.1041 -4.926 2.398e-06 * * *
cadmium -0.7432 0.1297 -5.729 6.212e-08 * * *
chromium 0.2695 0.1071 2.517 0.01301 *
mercury -0.7299 0.132 -5.529 1.588e-07 * * *
lead 1.43 0.1592 8.986 1.907e-15 * * *
## RMSE from cross-validation: 0.7065136
Variance Inflation Factors
  arsenic cadmium chromium mercury lead
VIF 1.56 2.08 1.75 1.43 2.7

Diversity/metals
## FULL MODEL
## Adjusted R2 is: 0.06
Fitting linear model: H ~ arsenic + cadmium + chromium + mercury + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.03115 0.07881 0.3952 0.6933
arsenic -0.115 0.1395 -0.8244 0.4112
cadmium -0.2287 0.1738 -1.316 0.1905
chromium -0.2563 0.1435 -1.786 0.0764
mercury 0.02354 0.1769 0.1331 0.8943
lead 0.2542 0.2133 1.192 0.2355
## RMSE from cross-validation: 0.9303706
Variance Inflation Factors
  arsenic cadmium chromium mercury lead
VIF 1.56 2.08 1.75 1.43 2.7

## REDUCED MODEL
## Adjusted R2 is: 0.07
Fitting linear model: H ~ chromium
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.0391 0.077 0.5078 0.6124
chromium -0.2812 0.08139 -3.455 0.0007288 * * *
## RMSE from cross-validation: 0.9157629
Variance Inflation Factors
  chromium
VIF 1

Evenness/metals
## FULL MODEL
## Adjusted R2 is: 0.03
Fitting linear model: J ~ arsenic + cadmium + chromium + mercury + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.07716 0.07843 0.9838 0.3269
arsenic 0.1621 0.1388 1.167 0.2451
cadmium 0.2536 0.173 1.466 0.145
chromium -0.238 0.1428 -1.666 0.09799
mercury 0.1657 0.176 0.9411 0.3483
lead -0.3184 0.2123 -1.5 0.1359
## RMSE from cross-validation: 0.9265784
Variance Inflation Factors
  arsenic cadmium chromium mercury lead
VIF 1.56 2.08 1.75 1.43 2.7

## REDUCED MODEL
## Adjusted R2 is: 0.02
Fitting linear model: J ~ chromium
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.05581 0.07726 0.7224 0.4713
chromium -0.1744 0.08166 -2.136 0.03445 *
## RMSE from cross-validation: 0.9269077
Variance Inflation Factors
  chromium
VIF 1

1 mm community
Richness/habitat
## FULL MODEL
## Adjusted R2 is: 0.25
Fitting linear model: S ~ depth + om + gravel + sand + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.04634 0.06256 -0.7407 0.4598
depth 0.2715 0.06482 4.188 4.307e-05 * * *
om -0.04458 0.1117 -0.3991 0.6903
gravel -0.1185 0.08561 -1.384 0.1679
sand -0.4367 0.1322 -3.304 0.001141 * *
clay -0.5032 0.1403 -3.586 0.0004272 * * *
## RMSE from cross-validation: 0.8934522
Variance Inflation Factors
  depth om gravel sand clay
VIF 1.06 1.81 1.03 2.1 1.45

## REDUCED MODEL
## Adjusted R2 is: 0.25
Fitting linear model: S ~ depth + sand + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.03361 0.06184 -0.5436 0.5874
depth 0.2678 0.06475 4.136 5.298e-05 * * *
sand -0.3869 0.07612 -5.083 8.825e-07 * * *
clay -0.4563 0.114 -4.001 9.001e-05 * * *
## RMSE from cross-validation: 0.8673699
Variance Inflation Factors
  depth sand clay
VIF 1.06 1.21 1.18

Density/habitat
## FULL MODEL
## Adjusted R2 is: 0.02
Fitting linear model: N ~ depth + om + gravel + sand + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.0211 0.07296 -0.2892 0.7727
depth -0.1713 0.07559 -2.266 0.02458 *
om -0.09757 0.1303 -0.749 0.4548
gravel 0.002914 0.09984 0.02919 0.9767
sand -0.29 0.1542 -1.881 0.06147
clay -0.3748 0.1636 -2.291 0.02309 *
## RMSE from cross-validation: 1.008745
Variance Inflation Factors
  depth om gravel sand clay
VIF 1.06 1.81 1.03 2.1 1.45

## REDUCED MODEL
## Adjusted R2 is: 0.03
Fitting linear model: N ~ depth + sand + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.01888 0.07186 -0.2627 0.7931
depth -0.1719 0.07525 -2.285 0.02341 *
sand -0.1969 0.08846 -2.226 0.02718 *
clay -0.3066 0.1325 -2.314 0.02174 *
## RMSE from cross-validation: 1.002598
Variance Inflation Factors
  depth sand clay
VIF 1.06 1.21 1.18

Diversity/habitat
## FULL MODEL
## Adjusted R2 is: 0.34
Fitting linear model: H ~ depth + om + gravel + sand + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.02566 0.05866 -0.4374 0.6623
depth 0.4303 0.06078 7.079 2.765e-11 * * *
om 0.06608 0.1047 0.631 0.5288
gravel -0.1077 0.08027 -1.341 0.1815
sand -0.2444 0.124 -1.971 0.05013
clay -0.2757 0.1316 -2.096 0.03742 *
## RMSE from cross-validation: 0.8438226
Variance Inflation Factors
  depth om gravel sand clay
VIF 1.06 1.81 1.03 2.1 1.45

## REDUCED MODEL
## Adjusted R2 is: 0.34
Fitting linear model: H ~ depth + gravel + sand + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) -0.02847 0.0584 -0.4875 0.6265
depth 0.4311 0.06067 7.105 2.348e-11 * * *
gravel -0.119 0.0781 -1.524 0.1292
sand -0.3082 0.07143 -4.315 2.564e-05 * * *
clay -0.3236 0.1073 -3.017 0.002901 * *
## RMSE from cross-validation: 0.8399062
Variance Inflation Factors
  depth gravel sand clay
VIF 1.06 1.01 1.21 1.18

Evenness/habitat
## FULL MODEL
## Adjusted R2 is: 0.08
Fitting linear model: J ~ depth + om + gravel + sand + clay
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.01566 0.06839 0.2289 0.8192
depth 0.2928 0.07086 4.133 5.383e-05 * * *
om 0.02954 0.1221 0.2419 0.8091
gravel 0.008159 0.09359 0.08718 0.9306
sand -0.04328 0.1445 -0.2995 0.7649
clay -0.02071 0.1534 -0.135 0.8927
## RMSE from cross-validation: 0.9552492
Variance Inflation Factors
  depth om gravel sand clay
VIF 1.06 1.81 1.03 2.1 1.45

## REDUCED MODEL
## Adjusted R2 is: 0.1
Fitting linear model: J ~ depth
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.01448 0.0664 0.2181 0.8276
depth 0.3126 0.06602 4.735 4.225e-06 * * *
## RMSE from cross-validation: 0.9358617
Variance Inflation Factors
  depth
VIF 1

Richness/metals
## FULL MODEL
## Adjusted R2 is: 0.07
Fitting linear model: S ~ arsenic + cadmium + chromium + mercury + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.06323 0.08253 0.7661 0.4451
arsenic -0.1466 0.1335 -1.098 0.2742
cadmium -0.5481 0.1606 -3.413 0.0008755 * * *
chromium 0.01217 0.1638 0.07431 0.9409
mercury -0.001392 0.1293 -0.01077 0.9914
lead 0.4503 0.2727 1.651 0.1013
## RMSE from cross-validation: 0.9298714
Variance Inflation Factors
  arsenic cadmium chromium mercury lead
VIF 1.65 1.85 1.89 1.48 3.17

## REDUCED MODEL
## Adjusted R2 is: 0.08
Fitting linear model: S ~ cadmium + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.06172 0.08196 0.7531 0.4529
cadmium -0.5082 0.149 -3.409 0.000881 * * *
lead 0.3104 0.1476 2.103 0.03751 *
## RMSE from cross-validation: 0.9219954
Variance Inflation Factors
  cadmium lead
VIF 1.73 1.73

Density/metals
## FULL MODEL
## Adjusted R2 is: -0.01
Fitting linear model: N ~ arsenic + cadmium + chromium + mercury + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.01452 0.09066 0.1602 0.873
arsenic -0.1569 0.1467 -1.07 0.287
cadmium -0.161 0.1764 -0.9126 0.3633
chromium -0.2011 0.18 -1.117 0.2661
mercury -0.1079 0.142 -0.7597 0.4489
lead 0.4704 0.2995 1.57 0.1189
## RMSE from cross-validation: 1.030112
Variance Inflation Factors
  arsenic cadmium chromium mercury lead
VIF 1.65 1.85 1.89 1.48 3.17

## REDUCED MODEL
## Adjusted R2 is: 0
Fitting linear model: N ~ 1
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.01849 0.08981 0.2058 0.8373
## RMSE from cross-validation: 1.011527

Quitting from lines 452-454 (C1_analyses_B.Rmd) Error in Qr$qr[p1, p1, drop = FALSE] : indice hors limites De plus : There were 33 warnings (use warnings() to see them)

Diversity/metals
## FULL MODEL
## Adjusted R2 is: 0.04
Fitting linear model: H ~ arsenic + cadmium + chromium + mercury + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.09394 0.08016 1.172 0.2435
arsenic -0.1326 0.1297 -1.023 0.3086
cadmium -0.4559 0.156 -2.923 0.004141 * *
chromium 0.006562 0.1591 0.04124 0.9672
mercury 0.03018 0.1255 0.2404 0.8104
lead 0.3954 0.2648 1.493 0.1381
## RMSE from cross-validation: 0.8995954
Variance Inflation Factors
  arsenic cadmium chromium mercury lead
VIF 1.65 1.85 1.89 1.48 3.17

## REDUCED MODEL
## Adjusted R2 is: 0.05
Fitting linear model: H ~ cadmium + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.09219 0.07962 1.158 0.2492
cadmium -0.4268 0.1448 -2.948 0.003831 * *
lead 0.2923 0.1434 2.038 0.04365 *
## RMSE from cross-validation: 0.8942355
Variance Inflation Factors
  cadmium lead
VIF 1.73 1.73

Evenness/metals
## FULL MODEL
## Adjusted R2 is: -0.04
Fitting linear model: J ~ arsenic + cadmium + chromium + mercury + lead
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.06202 0.07961 0.779 0.4375
arsenic -0.05693 0.1288 -0.4421 0.6592
cadmium -0.04309 0.1549 -0.2782 0.7813
chromium 0.05701 0.158 0.3607 0.7189
mercury 0.03799 0.1247 0.3047 0.7611
lead -0.004916 0.263 -0.01869 0.9851
## RMSE from cross-validation: 0.8928066
Variance Inflation Factors
  arsenic cadmium chromium mercury lead
VIF 1.65 1.85 1.89 1.48 3.17

## REDUCED MODEL
## Adjusted R2 is: 0
Fitting linear model: J ~ 1
  Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.06244 0.07799 0.8006 0.4249
## RMSE from cross-validation: 0.873627

Quitting from lines 476-478 (C1_analyses_B.Rmd) Error in Qr\(qr[p1, p1, drop = FALSE] : indice hors limites De plus : Warning messages: 1: In CVlm(data = lm_out\)model, form.lm = lm_out, m = 5, printit = F) :

As there is >1 explanatory variable, cross-validation predicted values for a fold are not a linear function of corresponding overall predicted values. Lines that are shown for the different folds are approximate

2: In CVlm(data = lm_out$model, form.lm = lm_out, m = 5, printit = F) :

As there is >1 explanatory variable, cross-validation predicted values for a fold are not a linear function of corresponding overall predicted values. Lines that are shown for the different folds are approximate

3: In CVlm(data = lm_out$model, form.lm = lm_out, m = 5, printit = F) :

As there is >1 explanatory variable, cross-validation predicted values for a fold are not a linear function of corresponding overall predicted values. Lines that are shown for the different folds are approximate

4: In CVlm(data = lm_out$model, form.lm = lm_out, m = 5, printit = F) :

As there is >1 explanatory variable, cross-validation predicted values for a fold are not a linear function of corresponding overall predicted values. Lines that are shown for the different folds are approximate

5. Multivariate regression

Independant variables are habitat parameters or heavy metal concentrations, dependant variables are species abundances for each community. Variables have been standardized by mean and standard-deviation, and outliers and correlated variables have been excluded.

This analysis has been done on PRIMER, with a DistLM to identify the variables that explain the most the community variability and with a dbRDA to plot the results.

0.5 mm community
Habitat

Variables selected by the DistLM procedure have a \(R^{2}\) of 0.27.

Metals

Variables selected by the DistLM procedure have a \(R^{2}\) of 0.18.

1 mm community
Habitat

Variables selected by the DistLM procedure have a \(R^{2}\) of 0.14.

Metals

Variables selected by the DistLM procedure have a \(R^{2}\) of 0.07.


| 📄